{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/deep-ehr-chronic-disease-prediction-using","title":"Deep EHR: Chronic Disease Prediction Using Medical Notes","arxiv_id":"1808.04928","date":"2018-08-15","proceeding":"Machine Learning for Health Care conference 2018 9","authors":["Jingshu Liu","Zachariah Zhang","Narges Razavian"],"abstract":"Early detection of preventable diseases is important for better disease\nmanagement, improved inter-ventions, and more efficient health-care resource\nallocation. Various machine learning approacheshave been developed to utilize\ninformation in Electronic Health Record (EHR) for this task. Majorityof\nprevious attempts, however, focus on structured fields and lose the vast amount\nof information inthe unstructured notes. In this work we propose a general\nmulti-task framework for disease onsetprediction that combines both free-text\nmedical notes and structured information. We compareperformance of different\ndeep learning architectures including CNN, LSTM and hierarchical models.In\ncontrast to traditional text-based prediction models, our approach does not\nrequire disease specificfeature engineering, and can handle negations and\nnumerical values that exist in the text. Ourresults on a cohort of about 1\nmillion patients show that models using text outperform modelsusing just\nstructured data, and that models capable of using numerical values and\nnegations in thetext, in addition to the raw text, further improve performance.\nAdditionally, we compare differentvisualization methods for medical\nprofessionals to interpret model predictions.","url_abs":"http://arxiv.org/abs/1808.04928v1","url_pdf":"http://arxiv.org/pdf/1808.04928v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"deep-ehr-chronic-disease-prediction-using","repo_url":"https://github.com/NYUMedML/DeepEHR","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":1,"framework":"pytorch","reach":{"status":"unanswered"}}],"tasks":[{"task_slug":"disease-prediction","task_name":"Disease Prediction"},{"task_slug":"management","task_name":"Management"},{"task_slug":"prediction","task_name":"Prediction"}],"methods":[{"method_slug":"lstm","method_name":"LSTM"},{"method_slug":"sigmoid-activation","method_name":"Sigmoid Activation"},{"method_slug":"tanh-activation","method_name":"Tanh Activation"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":"https://app.syntology.ai/?focus=1808.04928","mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}